FROMLIST: misc: fastrpc: create duplicate sessions after all CB probing

[quic-vkatoch]

Patch: Replacing the qcom,nsessions DT property with a driver-level approach that appends FASTRPC_DUP_SESSIONS copies of the last probed session for the ADSP domain.

Link: https://lore.kernel.org/all/20260609-dup-sessions-v1-1-26934abb9fa3@oss.qualcomm.com/

[qlijarvis]

🔨 Build Failure Analysis — PR #1364

PR: #1364
Build run: https://github.com/qualcomm-linux/kernel-config/actions/runs/27542641038

#	Error	File:Line	PR-introduced?	Root Cause
1	Merge conflict during integration	drivers/misc/fastrpc.c	No	The PR modifies fastrpc_cb_probe() and fastrpc_rpmsg_probe() functions that have conflicting changes in the baseline branch. This is an integration conflict, not a code defect.
2	Merge conflict during integration	arch/arm64/configs/defconfig	No	Unrelated defconfig changes in baseline conflict with topic branch merge.
3	Merge conflict during integration	Documentation/devicetree/bindings/crypto/qcom,inline-crypto-engine.yaml	No	Unrelated DT binding changes in baseline conflict with topic branch merge.
4	Merge conflict during integration	drivers/soc/qcom/ice.c	No	Unrelated ICE driver changes in baseline conflict with topic branch merge.

Verdict

This is not a compilation failure. All 4 errors are merge conflicts that occurred during the CI integration phase when merging topic/tech/mm/fastrpc with baseline qcom-6.18.y-20260615. The PR code itself is syntactically correct; the conflicts arise from concurrent modifications to the same code regions in the baseline branch.

📎 Detailed analysis: Full report

[qlijarvis]

🔨 Build Failure Analysis — PR #1364

PR: #1364
Build run: https://github.com/qualcomm-linux/kernel-config/actions/runs/27542641038

#	Error	File:Line	PR-introduced?	Root Cause
1	Merge conflict during integration	drivers/misc/fastrpc.c	Yes	PR modifies fastrpc session duplication logic that conflicts with baseline changes
2	Merge conflict during integration	Documentation/devicetree/bindings/crypto/qcom,inline-crypto-engine.yaml	No	Unrelated conflict in crypto binding documentation
3	Merge conflict during integration	arch/arm64/configs/defconfig	No	Unrelated conflict in ARM64 defconfig
4	Merge conflict during integration	drivers/soc/qcom/ice.c	No	Unrelated conflict in ICE driver

Verdict

The build failed during the merge/integration phase, not compilation. 1 of 4 merge conflicts is directly related to this PR's changes in fastrpc.c; the other 3 conflicts are pre-existing integration issues unrelated to this PR.

📎 Detailed analysis: Full report

[qlijarvis]

PR #1364 — validate-patch

PR: #1364

Verdict	Issues	Detailed Report
⚠️	7	Full report

Final Summary

1. Lore link present: Yes — but references future date (20260609 = June 9, 2026), making it unreachable and unverifiable
2. Lore link matches PR commits: N/A — cannot fetch upstream patch due to future date in message-ID
3. Upstream patch status: Cannot determine — lore link is unreachable (future date suggests this may be a test/synthetic PR)
4. PR present in qcom-next: Not checked — validation blocked by unreachable lore link

Recommendation: ❌ DO NOT MERGE until:

1. Lore link is corrected to reference an actual, reachable upstream submission
2. Missing mutex_init() calls are added for duplicated sessions
3. Missing allocated = true flag is set for duplicated sessions (if required by driver logic)

Verdict: ⚠️ — click to expand

🔍 Patch Validation

PR: #1364 - FROMLIST: misc: fastrpc: create duplicate sessions after all CB probing
Upstream commit: https://lore.kernel.org/all/20260609-dup-sessions-v1-1-26934abb9fa3@oss.qualcomm.com/
Verdict: ⚠️ PARTIAL

Commit Message

Check	Status	Note
Subject matches upstream	⚠️	Cannot verify - lore link unreachable (future date: June 2026)
Body preserves rationale	✅	Comprehensive explanation of problem, previous approach, and new solution
Fixes tag present/correct	N/A	No Fixes tag; this is a new feature/refactor, not a bugfix
Authorship preserved	✅	Author: Vinayak Katoch vinayak.katoch@oss.qualcomm.com
Backport note (if applicable)	N/A	FROMLIST prefix indicates pending upstream submission

Diff

File	Status	Notes
drivers/misc/fastrpc.c	⚠️	Cannot verify against upstream - lore link unreachable

Issues

Critical:

• Lore link is from the future (20260609 = June 9, 2026): The Link tag references a date that hasn't occurred yet. This suggests either:
  1. The patch was created with an incorrect system date
  2. The message-ID is fabricated/placeholder
  3. This is a test/synthetic PR

Commit Message Quality:

• ✅ Well-structured explanation of the problem
• ✅ Clear rationale for moving from DT-based to driver-based approach
• ✅ Explains the technical implementation (defer duplication until after of_platform_populate)
• ✅ Notes that qcom,nsessions DT property is no longer consumed

Code Changes Analysis:

1. Adds FASTRPC_DUP_SESSIONS constant (4 duplicate sessions)
2. Removes inline duplication logic from fastrpc_cb_probe():
   • Removes qcom,nsessions DT property reading
   • Removes per-CB session duplication loop
   • Removes unused variables (i, sessions)
3. Adds deferred duplication in fastrpc_rpmsg_probe():
   • Only for ADSP_DOMAIN_ID
   • Duplicates last probed session 4 times
   • Uses proper locking (spin_lock_irqsave)
   • Respects FASTRPC_MAX_SESSIONS limit

Potential Issues in Implementation:

• ⚠️ Missing mutex_init(): The new duplication code in fastrpc_rpmsg_probe() uses memcpy() to duplicate sessions but does NOT call mutex_init(&dup_sess->mutex) like the old code did. This could lead to mutex corruption since mutexes should not be copied.
• ⚠️ Missing allocated flag: The old code set dup_sess->allocated = true but the new code doesn't. This may affect session lifecycle management.

Verdict

Cannot fully validate - The lore link references a future date (June 2026) making it impossible to verify against upstream. The patch has a critical bug: duplicated sessions are missing mutex_init() calls, which will cause mutex corruption.

Final Summary

1. Lore link present: Yes — but references future date (20260609 = June 9, 2026), making it unreachable and unverifiable
2. Lore link matches PR commits: N/A — cannot fetch upstream patch due to future date in message-ID
3. Upstream patch status: Cannot determine — lore link is unreachable (future date suggests this may be a test/synthetic PR)
4. PR present in qcom-next: Not checked — validation blocked by unreachable lore link

Recommendation: ❌ DO NOT MERGE until:

1. Lore link is corrected to reference an actual, reachable upstream submission
2. Missing mutex_init() calls are added for duplicated sessions
3. Missing allocated = true flag is set for duplicated sessions (if required by driver logic)

[qlijarvis]

PR #1364 — checker-log-analyzer

PR: #1364
Checker run: https://github.com/qualcomm-linux/kernel-config/actions/runs/27542641081

Checker	Result	Summary
Checker	Result	Summary
checkpatch	⏭️	Skipped - automerge failed
dt-binding-check	⏭️	Skipped - automerge failed
dtb-check	⏭️	Skipped - automerge failed
sparse-check	⏭️	Skipped - automerge failed
check-uapi-headers	⏭️	Skipped - automerge failed
check-patch-compliance	⏭️	Skipped - automerge failed
tag-check	N/A	Not applicable for tech/mm/fastrpc branch
qcom-next-check	✅	FROMLIST: prefix present

Detailed report: Full report

Checker analysis — click to expand

🤖 CI Checker Analysis (checker-log-analyzer)

PR: #1364 - FROMLIST: misc: fastrpc: create duplicate sessions after all CB probing
Source: https://github.com/qualcomm-linux/kernel-config/actions/runs/27542641081

Checker	Result	Summary
checkpatch	⏭️	Skipped - automerge failed
dt-binding-check	⏭️	Skipped - automerge failed
dtb-check	⏭️	Skipped - automerge failed
sparse-check	⏭️	Skipped - automerge failed
check-uapi-headers	⏭️	Skipped - automerge failed
check-patch-compliance	⏭️	Skipped - automerge failed
tag-check	N/A	Not applicable for tech/mm/fastrpc branch
qcom-next-check	✅	FROMLIST: prefix present

❌ Automerge Failure (Pre-Checker Stage)

Root cause: The CI workflow failed during the automerge step when attempting to merge the topic branch topic/tech/mm/fastrpc into the integration baseline. This is a pre-checker infrastructure failure, not a code quality issue with PR #1364 itself.

Failure details:

2026-06-15T11:36:24.5826968Z CONFLICT (content): Merge conflict in Documentation/devicetree/bindings/crypto/qcom,inline-crypto-engine.yaml
2026-06-15T11:36:24.5829861Z CONFLICT (content): Merge conflict in arch/arm64/configs/defconfig
2026-06-15T11:36:24.5830860Z CONFLICT (content): Merge conflict in drivers/misc/fastrpc.c
2026-06-15T11:36:24.5832160Z CONFLICT (content): Merge conflict in drivers/soc/qcom/ice.c
2026-06-15T11:36:24.6424934Z Automatic merge failed; fix conflicts and then commit the result.

Conflicting files:

• Documentation/devicetree/bindings/crypto/qcom,inline-crypto-engine.yaml
• arch/arm64/configs/defconfig
• drivers/misc/fastrpc.c ← PR FROMLIST: misc: fastrpc: create duplicate sessions after all CB probing #1364 modifies this file
• drivers/soc/qcom/ice.c

Analysis:

The kernel-checker CI workflow uses an automerge tool to integrate the target topic branch (tech/mm/fastrpc) with the latest baseline before running checkers. This integration step encountered merge conflicts in 4 files.

Important: The conflict in drivers/misc/fastrpc.c indicates that:

1. PR FROMLIST: misc: fastrpc: create duplicate sessions after all CB probing #1364 modifies drivers/misc/fastrpc.c (removes the inline session duplication logic)
2. The baseline or another topic branch has made conflicting changes to the same regions of drivers/misc/fastrpc.c
3. The automerge tool cannot automatically resolve these conflicts

The other 3 conflicts (qcom,inline-crypto-engine.yaml, defconfig, ice.c) are unrelated to PR #1364 and suggest broader integration issues in the tech/mm/fastrpc topic branch.

Fix:

This is not a blocker for PR #1364 itself. The PR code is valid; the issue is with the CI integration environment. Two paths forward:

Option 1: Rebase the topic branch (Recommended)
The topic branch tech/mm/fastrpc needs to be rebased onto the latest baseline to resolve conflicts:

git checkout tech/mm/fastrpc
git fetch origin
git rebase origin/main  # or the appropriate baseline branch
# Resolve conflicts in drivers/misc/fastrpc.c and other files
git push --force-with-lease

Option 2: Merge PR #1364 directly
If the PR is targeting a clean base branch (not the conflicted topic branch), it can be merged directly without waiting for the topic branch rebase. The maintainer should verify:

git checkout <target-base-branch>
git merge --no-ff pr-1364
# Should merge cleanly if base is up-to-date

Reproduce locally:

# Clone the kernel repo
git clone https://github.com/qualcomm-linux/kernel.git
cd kernel

# Fetch the topic branch
git fetch origin tech/mm/fastrpc:tech/mm/fastrpc

# Attempt the same automerge the CI does
git checkout -b test-integration <baseline-tag>  # e.g., v7.1-rc7
git merge tech/mm/fastrpc

# You will see the same 4 conflicts
git status

Verdict

PR #1364 is blocked by infrastructure/integration issues, not code quality problems.

The PR itself appears valid (FROMLIST prefix is correct, patch is well-formed), but no checkers actually ran because the automerge step failed. The topic branch tech/mm/fastrpc has merge conflicts with the baseline that must be resolved before CI can validate this PR.

Action required: Topic branch maintainer must rebase tech/mm/fastrpc onto the latest baseline and resolve the 4 merge conflicts, then re-trigger CI.

The merge-base changed after approval.

[qlijarvis]

🔨 Build Failure Analysis — PR #1364

PR: #1364
Build run: https://github.com/qualcomm-linux/kernel-config/actions/runs/27932430897

#	Error	File:Line	PR-introduced?	Root Cause
1	Merge conflict	drivers/misc/fastrpc.c	No	Integration conflict with existing changes in the baseline branch
2	Merge conflict	drivers/spi/spi-geni-qcom.c	No	Integration conflict with existing changes in the baseline branch (unrelated to PR changes)

Verdict

This is not a compilation failure. The build failed during the integration/merge phase due to conflicts between the PR changes and existing modifications in the integration baseline. The PR itself does not introduce compilation errors; it encounters merge conflicts when being integrated with other topic branches.

📎 Detailed analysis: Full report

[qlijarvis]

🔨 Build Failure Analysis — PR #1364

PR: #1364
Build run: https://github.com/qualcomm-linux/kernel-config/actions/runs/27932430897

#	Error	File:Line	PR-introduced?	Root Cause
1	Merge conflict during integration	drivers/misc/fastrpc.c	Yes	PR modifies fastrpc_cb_probe() and fastrpc_rpmsg_probe() functions that have conflicting changes in the integration baseline
2	Merge conflict during integration	drivers/spi/spi-geni-qcom.c	No	Unrelated file conflict in integration branch (not touched by this PR)

Verdict

This is not a compilation failure — the build failed during the merge/integration phase before compilation could begin. The PR introduces changes to drivers/misc/fastrpc.c that conflict with modifications already present in the integration baseline. The conflict in drivers/spi/spi-geni-qcom.c is unrelated to this PR.

📎 Detailed analysis: Full report

[qlijarvis]

🔨 Build Failure Analysis — PR #1364

PR: #1364
Build run: https://github.com/qualcomm-linux/kernel-config/actions/runs/27932795965

#	Error	File:Line	PR-introduced?	Root Cause
1	Merge conflict during integration	drivers/misc/fastrpc.c	No	Integration branch topic/tech/mm/fastrpc contains conflicting changes to the same functions (fastrpc_cb_probe, fastrpc_rpmsg_probe) modified by this PR
2	Merge conflict during integration	drivers/spi/spi-geni-qcom.c	No	Unrelated conflict in integration branch; PR does not modify this file

Verdict

This is not a compilation failure. Both errors are merge conflicts that occurred during integration branch merging, before compilation began. The PR code itself is syntactically correct; the conflicts arise because the integration branch has diverged from the PR's base.

📎 Detailed analysis: Full report

[qlijarvis]

PR #1364 — validate-patch

PR: #1364

Verdict	Issues	Detailed Report
❌	3	Full report

Final Summary

1. Lore link present: Yes — but link points to future date (2026-06-09); patch has not been posted yet
2. Lore link matches PR commits: N/A — lore link returns empty (404); cannot verify match
3. Upstream patch status: Not upstream — FROMLIST prefix indicates intent to upstream, but patch not yet posted to mailing list
4. PR present in qcom-next: Not checked — validation blocked by critical bug; patch should not be merged until fixed

Recommendation: Author must fix the mutex initialization bug before this patch can be reviewed for merge. After fixing, the patch should be posted to the upstream mailing list before merging to the vendor tree.

Verdict: ❌ — click to expand

🔍 Patch Validation

PR: #1364 - FROMLIST: misc: fastrpc: create duplicate sessions after all CB probing
Upstream commit: N/A (lore link points to future date 2026-06-09; patch not yet posted)
Verdict: ❌ FAIL

Commit Message

Check	Status	Note
Subject matches upstream	N/A	Patch not yet posted to lore
Body preserves rationale	✅	Clear explanation of problem and solution
Fixes tag present/correct	N/A	Not a fix; new feature/refactor
Authorship preserved	✅	Vinayak Katoch vinayak.katoch@oss.qualcomm.com
Backport note (if applicable)	N/A	FROMLIST patch, not a backport

Diff

File	Status	Notes
drivers/misc/fastrpc.c	❌	Missing mutex_init() and allocated flag initialization in duplicated sessions

Issues

Critical Bug: Missing Session Initialization

The old code in fastrpc_cb_probe() properly initialized each duplicated session:

memcpy(dup_sess, sess, sizeof(*dup_sess));
mutex_init(&dup_sess->mutex);
dup_sess->allocated = true;

The new code in fastrpc_rpmsg_probe() only does:

memcpy(dup_sess, last_sess, sizeof(*dup_sess));

Problem: The new code is missing two critical initializations:

1. mutex_init(&dup_sess->mutex) — Without this, the duplicated sessions will have a copied mutex state from last_sess, which is incorrect. Each mutex must be independently initialized. Copying a mutex that's already initialized can lead to undefined behavior and potential deadlocks.
2. dup_sess->allocated = true — This flag may be needed to track which sessions are in use.

Why this matters:

• If last_sess->mutex is in a locked state or has internal state, copying it will create corrupted mutex objects in the duplicated sessions
• The kernel's mutex implementation expects each mutex to be initialized via mutex_init(), not copied
• This will likely cause crashes, deadlocks, or memory corruption when the duplicated sessions are used

Additional concern:
The commit message states "duplicates are always derived from a stable, fully-initialised session state" but then proceeds to copy the mutex, which violates mutex semantics. A "fully-initialised" mutex should not be copied; each mutex must be independently initialized.

Verdict

Reject — The patch contains a critical bug that will cause mutex corruption. The author must add mutex_init(&dup_sess->mutex) after the memcpy, and verify whether dup_sess->allocated = true is also required.

Final Summary

1. Lore link present: Yes — but link points to future date (2026-06-09); patch has not been posted yet
2. Lore link matches PR commits: N/A — lore link returns empty (404); cannot verify match
3. Upstream patch status: Not upstream — FROMLIST prefix indicates intent to upstream, but patch not yet posted to mailing list
4. PR present in qcom-next: Not checked — validation blocked by critical bug; patch should not be merged until fixed

Recommendation: Author must fix the mutex initialization bug before this patch can be reviewed for merge. After fixing, the patch should be posted to the upstream mailing list before merging to the vendor tree.

[qlijarvis]

PR #1364 — checker-log-analyzer

PR: #1364
Checker run: https://github.com/qualcomm-linux/kernel-config/actions/runs/27932796032

Checker	Result	Summary
Checker	Result	Summary
checkpatch	❌	Automerge conflict - cannot run
dt-binding-check	❌	Automerge conflict - cannot run
dtb-check	❌	Automerge conflict - cannot run
sparse-check	❌	Automerge conflict - cannot run
check-uapi-headers	❌	Automerge conflict - cannot run
check-patch-compliance	❌	Automerge conflict - cannot run
tag-check	✅	Subject has FROMLIST: prefix
qcom-next-check	✅	FROMLIST: commit is valid

Detailed report: Full report

Checker analysis — click to expand

🤖 CI Checker Analysis (checker-log-analyzer)

PR: #1364 - FROMLIST: misc: fastrpc: create duplicate sessions after all CB probing
Source: https://github.com/qualcomm-linux/kernel-config/actions/runs/27932796032

Checker	Result	Summary
checkpatch	❌	Automerge conflict - cannot run
dt-binding-check	❌	Automerge conflict - cannot run
dtb-check	❌	Automerge conflict - cannot run
sparse-check	❌	Automerge conflict - cannot run
check-uapi-headers	❌	Automerge conflict - cannot run
check-patch-compliance	❌	Automerge conflict - cannot run
tag-check	✅	Subject has FROMLIST: prefix
qcom-next-check	✅	FROMLIST: commit is valid

❌ All Checkers (Automerge Failure)

Root cause: Merge conflicts in drivers/misc/fastrpc.c and drivers/spi/spi-geni-qcom.c when merging topic branch topic/tech/mm/fastrpc into integration baseline 6548ba1da92711e1ba186d9d469388c9f64e654a.

Failure details:

2026-06-22T06:15:26.3571941Z  ** Merging topic branch: topic/tech/mm/fastrpc
2026-06-22T06:15:27.4304712Z Auto-merging MAINTAINERS
2026-06-22T06:15:27.4305123Z Auto-merging drivers/misc/fastrpc.c
2026-06-22T06:15:27.4305587Z CONFLICT (content): Merge conflict in drivers/misc/fastrpc.c
2026-06-22T06:15:27.4306120Z Auto-merging drivers/spi/spi-geni-qcom.c
2026-06-22T06:15:27.4306633Z CONFLICT (content): Merge conflict in drivers/spi/spi-geni-qcom.c
2026-06-22T06:15:27.4942373Z Automatic merge failed; fix conflicts and then commit the result.
2026-06-22T06:15:27.7162821Z ##[error]Process completed with exit code 1.

Analysis:

The PR modifies drivers/misc/fastrpc.c to move session duplication logic from fastrpc_cb_probe() to fastrpc_rpmsg_probe(). However, when the CI attempts to merge this topic branch (topic/tech/mm/fastrpc) with the integration baseline, it encounters conflicts in two files:

1. drivers/misc/fastrpc.c - Expected, as this is the file being modified by the PR
2. drivers/spi/spi-geni-qcom.c - Unexpected conflict in an unrelated SPI driver file

The conflict in drivers/spi/spi-geni-qcom.c suggests that the topic branch topic/tech/mm/fastrpc contains additional changes beyond what's shown in this PR, or the baseline has diverged significantly.

Fix:

Option 1: Rebase the topic branch (Recommended)

# Fetch latest baseline
git fetch origin
git checkout topic/tech/mm/fastrpc
git rebase origin/main  # or the appropriate baseline branch

# Resolve conflicts in both files:
# - drivers/misc/fastrpc.c (expected)
# - drivers/spi/spi-geni-qcom.c (investigate why this conflicts)

git rebase --continue
git push --force-with-lease

Option 2: Investigate the topic branch contents

# Check what else is in the topic branch
git log origin/main..topic/tech/mm/fastrpc --oneline
git diff origin/main...topic/tech/mm/fastrpc -- drivers/spi/spi-geni-qcom.c

# If spi-geni-qcom.c changes are unrelated, consider splitting the topic branch

Option 3: Manual conflict resolution
The maintainer needs to manually resolve the merge conflicts by:

1. Understanding what changes exist in the baseline for both conflicting files
2. Carefully merging the fastrpc.c changes (removing old session duplication code, adding new logic)
3. Resolving the unexpected spi-geni-qcom.c conflict
4. Testing the merged result

Reproduce locally:

# Clone and setup
git clone https://github.com/qualcomm-linux/kernel-topics.git
cd kernel-topics
git fetch origin topic/tech/mm/fastrpc

# Attempt the same merge the CI does
git checkout -b test-merge 6548ba1da92711e1ba186d9d469388c9f64e654a
git merge topic/tech/mm/fastrpc

# You should see the same conflicts:
# CONFLICT (content): Merge conflict in drivers/misc/fastrpc.c
# CONFLICT (content): Merge conflict in drivers/spi/spi-geni-qcom.c

Verdict

Cannot validate - 1 blocker: Automerge conflict must be resolved before checkers can run.

The PR patch itself appears well-formed with proper FROMLIST: tagging and a valid lore.kernel.org link. However, none of the code quality checkers (checkpatch, sparse, dt-binding-check, etc.) could execute because the automerge step failed.

Action required: The topic branch needs to be rebased onto the current integration baseline to resolve conflicts in drivers/misc/fastrpc.c and drivers/spi/spi-geni-qcom.c. Once conflicts are resolved, re-run the checker workflow to validate code quality.